Newsgroups: comp.robotics
Path: brunix!sgiblab!darwin.sura.net!emory!news-feed-1.peachnet.edu!gatech!howland.reston.ans.net!cs.utexas.edu!utnut!torn!newshub.ccs.yorku.ca!newshub.ariel.cs.yorku.ca!cs911225
From: cs911225@ariel.cs.yorku.ca (KEN E WILLMOTT)
Subject: Re: IR receivers
Message-ID: <CJ83IL.Isn@ariel.cs.yorku.ca>
Sender: news@ariel.cs.yorku.ca
Organization: York University, Dept. of Computer Science
References: <2gcf1e$i2c@handler.Eng.Sun.COM>
Date: Thu, 6 Jan 1994 19:30:20 GMT
Lines: 51

cmcmanis@pepper.Eng.Sun.COM (Chuck McManis)
says:

>So I've been playing around with some Sharp IR receiver modules, in particular
>the GP1U52X and the ISU160. The latter is available from Pure Unobtanium and
>is, in my opinion a significant improvement. (no metal box, 1/6th the volume,
...

Wow, I'll have to keep my eyes open for that part.

>However, I figured I rig up a 38Khz source (the 160's are tuned to 38Khz
>and _not_ 40Khz) using a 555 and a simple 50% duty cycle circuit I had in 
>a book on the 555. Since the 555 is loath to generate 50% duty cycles, I
...

Why, oh, why are people still using this 70's component?

...[efforts to make 555 circuit work deleted]...

>Hmmm, I say, could the IR sensor be derated by jitter in the modulation
>carrier? Well lets see, hook up the old HP signal generator and feed the
>IR LED 38,000.0 Khz exactly (through a buffer amp of course). Whoa! I can
>see for miles and miles. Sigh. Bottom line, it is seriously important to
>make the carrier as accurate as possible. 

I assume you mean phase jitter. I don't think that was your problem,
though. You never actually observed any, you just made an assumption
that that was the problem. I'll bet that if there was phase jitter
in a given signal, it would degrade performance, though.


>So I now know of at least three variables that contribute
>to the effectiveness of the IR sensor/LED pair's sensitivity:
>	1) "Jitter" in the input signal. Tolerance, low.
>	2) "Frequency" in the input signal, both sensors have 4Khz band
>	    pass filters centered at 38 and 40Khz respectively. 
>	    Tolerance moderate.
>	3) "Duty cycles" of the carrier. Both sensors respond best to a
>	   50% duty cycle square wave. Tolerence medium to low. Duty cycles
>	   are sometimes reduced to allow higher current through the LED.
>	   I've come to the conclusion it is cheaper/better to simply run
>	   multiple LEDs rather than boost the current on one.
>
>--Chuck

This is right on. However, the GP1U52X (and others like it) operate
at 41.3 Khz, not 40 Khz. That figure comes from the fact that some
transmitters used a 455Khz crystal resonator as a timebase, which
when divided by 11, gives 41.364 Khz.

	-Ken Willmott
